Five Claude skills an everyday business operator can paste, run, and use by Monday morning
Maria runs a bookkeeping practice in Hialeah. She has eleven clients, one assistant, and a phone that does not stop. Every Sunday night she sits down to write the week's client update emails, pull numbers from three different spreadsheets, check her appointment calendar, and draft follow-up messages for the clients who have not responded to invoices. It takes her four hours. She has been doing it this way for six years. She is not lazy. She is not slow. She just never had a faster way.
The promise of AI reached Maria the same way it reaches most people. A podcast. A LinkedIn post. A nephew at Thanksgiving who said she should be using ChatGPT. She downloaded an app, typed a question, got a wall of text that did not sound like her, and closed the tab. That was fourteen months ago. The Sunday nights kept coming. The four hours did not shrink. The gap between what AI promised and what Maria actually experienced stayed exactly where it was.
This book does not close that gap with theory. It closes it with five specific things you build, in order, over a single weekend. Each one is a working prompt system you paste into Claude, adapt to your business, and use on Monday. By the time you finish, you will have a client communication engine, a meeting intelligence tool, a proposal builder, an onboarding sequence, and a weekly planning system. Maria has all five now. Her Sunday nights run forty minutes.
The inbox does not care about your schedule. A client sends a message at 11 PM on a Tuesday, and by Wednesday morning it is already a day old and you have not answered it. Not because you are ignoring them. Because you opened it, read it, could not figure out the right tone, and closed it. The message is still there. The client is still waiting.
Client communication failure is almost never about what to say. It is about the four decisions that happen before the first word gets written: what kind of message is this, what tone does the relationship require right now, what do I actually want them to do when they finish reading, and how do I close without sounding like I am either begging or dismissing. Those four decisions under time pressure, on a day when twelve other things are already moving, is why the draft sits unsent.
The Client Communication Prompt strips those four decisions down to three questions it asks before writing anything. You answer. It writes. The draft that comes back sounds like you because you told it how you sound.
Here is the prompt:
You are my client communication assistant. My name is [your name] and I run [your business type]. My communication style is direct and warm. I do not use filler phrases. I do not open with "I hope this finds you well." I close with one specific next step, never a vague offer to connect. When I bring you a client situation, ask me these three questions before writing: What is the relationship stage — new, ongoing, or strained? What do I want the client to do after reading this? Is there a deadline? After I answer, write one email. Not options. One email. Under 150 words unless I tell you the situation needs more.
The first time you run this, bring it the message that has been sitting in your drafts the longest. Answer the three questions. Read what comes back. That draft will be shorter than what you would have written, cleaner than what you would have written, and it will close on something specific. Send it.
Here's What We Built. We built this exact prompt structure into our own publishing operation. The client communication system you just installed is one of the five running in our workflow right now. Not a demo. A live process that produces real output every week.
The meeting ends. Everyone nods. The person who was taking notes on their laptop sends a summary two days later that captures about sixty percent of what was decided and none of what was assigned. The other forty percent lives in three different people's memories, which will produce three slightly different versions of what happened by the following week. By the time the next meeting starts, the first ten minutes go to reconstructing the previous one. Nothing moves. The same conversation happens again with different words.
Meeting intelligence is not a note-taking problem. It is a structure problem. Notes are a stream. Decisions, owners, deadlines, and open questions are a document. The stream never becomes the document on its own. Someone has to convert it. Nobody ever has time to convert it. The meeting recedes into organizational memory and the next one starts with the same gaps.
The Meeting Intelligence Prompt converts the stream to the document in under two minutes. You paste your raw notes, bullet points, voice memo transcript, anything you have, and it returns four things: every decision made, every task assigned with an owner, every deadline named, and every question raised that was not answered. It flags missing owners and missing deadlines so nothing disappears into ambiguity.
Here is the prompt:
You are my meeting documentation assistant. I will paste raw notes from a business meeting. Return exactly four sections with no additional commentary. DECISIONS: every specific decision made, written as a completed statement. OWNERS: every task assigned and the name of the person who owns it — flag any unassigned tasks. DEADLINES: every deadline mentioned, tied to its task and owner — flag any missing deadlines. OPEN QUESTIONS: every question raised that was not resolved. After producing the document, ask me: should I draft a recap email to send to attendees?
Run this after every meeting. Not the important ones. Every one. The value compounds. You will start to notice patterns in which meetings produce decisions and which ones produce conversations. That is information you can act on.
Here's What We Built. We run this meeting intelligence process on every internal review. Raw notes go in. A structured decision log comes out. We did not build it for the book and then stop. It is still running.
He needed the quote by Wednesday. You said you would get it to him. Wednesday came and you were handling something else. Thursday you remembered but it felt too late to call without an explanation, and a quote without a conversation felt cold. By Monday it had been a week. The moment had passed. The job went somewhere else.
Proposal delay is not a discipline problem. It is a friction problem. To write a proper quote, you have to reconstruct the conversation from memory, translate the client's language into your pricing language, figure out how to present the number without apologizing for it, write a scope that covers you without reading like a legal document, and format all of it into something that looks like it came from a real business. That is five tasks compressed into one. Under any amount of pressure, five tasks compress into zero.
The Proposal Prompt takes four of those five tasks off your plate. The only thing it cannot do is know your numbers. You bring the numbers. It builds the document.
Here is the prompt:
You are my proposal assistant. When I give you a client situation, ask me: the client's name and business, the service they are requesting, my price for this, the timeline I am committing to, and any exclusions they need to understand. After I answer, produce a proposal with five sections: a one-paragraph plain-language project summary, a clear scope of work including what is not covered, the investment stated directly without hedging, the timeline, and the next step for the client to confirm and pay. Write in direct professional prose. No jargon. Do not use the word deliverables unless I use it first. After the proposal, ask me: do you want a follow-up message to send if you do not hear back in three days?
Before the next chapter, open this prompt and bring it one proposal you need to send this week. Answer its questions. Send what comes back.
Here's What We Built. The proposal logic in this chapter is the same logic behind every offer we put in front of a buyer. We do not hedge the price. We state the scope. We name the next step. That is not a philosophy. It is the template we actually use.
She said yes on the call. She was ready. You sent the intake form two hours later and felt good about it. Four days passed. You sent a reminder. Nothing. Two weeks later she emailed asking if you were still available. You were. But the energy was gone and both of you knew it.
Onboarding dropout is the most expensive problem a service business has because it is invisible on the revenue line. The client said yes. They counted. Then they did not show up. The loss never appears as a lost sale because the sale technically closed. It appears as a ghost client, a delayed start, an awkward re-engagement conversation that takes twice as long as the original close.
The dropout happens in the gap between yes and first delivery. The intake form was the gap. It looked like a form. It read like homework. The client opened it, saw twelve fields with unclear instructions, decided to do it later, and later never came.
The Onboarding Sequence Prompt replaces the form with a conversation. You paste it into Claude, give it your business details, and it runs a guided intake that feels like talking to someone who is actually interested in the client's problem. It asks one question at a time. It acknowledges answers before moving forward. When it finishes, it hands you a clean brief in the client's own words. That brief is worth more than any intake form because the language in it is the language you use when you write back to them.
Here is the prompt:
You are my client onboarding assistant. My business is [describe your business in one sentence]. When a new client starts a conversation with you, welcome them and gather the following through natural conversation, one question at a time: their full name and preferred name, their business and what they do, the specific problem they need solved, any previous attempts to solve it and why those did not work, their timeline and hard deadlines, their communication preference, and their budget range if they are comfortable sharing. When you have everything, say: I have everything I need to get you started. I will share this with [your name] and you will hear from us within [your timeframe]. Then give me a clean summary brief in under two minutes of reading.
The check-in question is the part most people skip. In week two, ask the new client: on a scale of one to ten, how confident do you feel about where we are right now? If the answer is below eight, you want to know that in week two. Not in week six.
Before the next chapter, run this prompt for your most recently signed client and compare it to what you actually sent them in their first thirty days.
Here's What We Built. We built an onboarding sequence using this exact structure and ran it on our own clients. The check-in question in week two has surfaced problems early every time we have used it. That is not a theory. That is what the record shows.
Sunday at 9 PM is the worst time to figure out what the week is. By then the brain has already started protecting itself from Monday. The resistance is not laziness. It is self-defense. The mind knows the week is coming and it does not have a map, so it resists looking at the territory. The result is that Monday morning arrives with no plan, the first fire of the day sets the agenda, and by Thursday you are wondering why the important work never got done.
The Weekly Planning Prompt does not make the week easier. It makes it visible.
The prompt runs on Sunday and it takes twelve minutes. The output is a working document: a priority stack, a daily time allocation, a list of decisions that need to be made before Wednesday, and a short list of communications that should go out before anything else touches your calendar. It is not a to-do list. A to-do list is a collection of tasks without a sequence. This is a sequence.
Here is the prompt:
I am planning my work week. Today is Sunday. My business is [describe your business]. Here is what is on my plate this week: [list your active clients, projects, and deadlines]. Here is what did not get done last week that carries forward: [list the carryovers]. Here is one thing that has to be done by Friday no matter what: [name it]. Build me a weekly plan. Start with a priority stack of no more than five items. Then allocate each day of the work week with one primary focus and two secondary tasks. Then give me a list of communications I should send Monday morning to clear the path. Then give me a list of decisions I need to make before Wednesday so nothing stalls. Keep the language direct and practical.
Run this prompt every Sunday. Not sometimes. Every Sunday. The value is not in any single week's plan. The value is in the pattern that builds when you do it forty weeks in a row. You start to see which clients always generate carryovers. You start to see which days lose their primary focus by 11 AM. You start to see the structure of how your work actually moves versus how you thought it moved.
The decisions list in the output is the part most people skip the first time. Do not skip it. A stalled project is almost always a delayed decision. The prompt surfaces the decision before the stall happens. Stalls cost more time than decisions do. Even a bad decision moves faster than no decision at all.
The Monday morning communication stack is the last piece. These are the messages you send before you open anything else. Not replies. Not responses to what came in over the weekend. The messages you initiate. The client who needs a status update before they call you to ask for one. The vendor who needs a confirmation before their timeline slips. The assistant who needs the week's priorities before they start building their own plan. Send those first. Then open your inbox. The week runs differently when you start it instead of responding to it.
Before you close this chapter, do this one thing: run the weekly planning prompt this Sunday, and do not skip the decisions list.
Here's What We Built. We run this planning prompt every week. The decisions list is the part that changed how we operate — not the priority stack, not the daily anchors. Surfacing decisions before they become stalls is the move. We figured that out by running the prompt, not by writing about it.
You finished the book. That means you have five working systems sitting in Claude Projects right now — or you will by Monday if you do the assignments. Not five ideas. Five things that run.
Here is what the record shows: a bookkeeper in Hialeah cut her Sunday nights from four hours to forty minutes. Not because AI is magic. Because she finally gave it a real job. That is what this book was for.
We built all five of these systems. We use all five. The output you are reading came from a pipeline that runs on the same logic. The proof is not a case study on the back cover. It is the book itself.
Everything we have built from this process is at aixsweb.com. Come find what we built. Then build something of your own.
aixsweb.com
---
Practical guides for small business owners building online the right way.
No spam. Unsubscribe anytime.
You're in. Next guide coming your way.